Investigation of the Dayton IR 75 sign truss failure of 9/11/06.

ODOT Agreement Number 14237Based upon a combination of in-situ field monitoring of traffic-induced bridgevibrations at the location of the failed sign support truss, finite element simulation of theexpected dynamic response of the original truss in such an environment, the length ofservice of the truss at the time of its failure, the volume of truck traffic on the bridge inquestion during that time of service, and metallurgical examination of the failedcomponents of the sign truss, the conclusion of the investigation is that extremely high-cycle fatigue of the chord/web diagonal welded connection was the cause of the trussfailure. The in-service effective stress range of the AASHTO Category ET connection inquestion was most likely below the currently specified AASHTO constant amplitudefatigue limit (CAFL) for the detail, but the enormous quantity of response cycles(approaching or even exceeding 1 billion cycles) such a bridge mounted sign on aheavily traveled route accumulates over a service lifetime of 30 or 40 years, exceedsanything currently considered in the design codes for such structures. The implicationsof this fact on the current inventory of such structures in similar installations and of thesame age range, i.e., installed with the original interstate routes, largely in the 1960\u2019s,are obvious.The one somewhat puzzling aspect of the lower chord fracture of this structure,namely that the fracture passed near the weld toe but not actually contacting the weldtoe, does not appear to be particularly significant. The fracture did indeed pass close tothe weld and certainly within the heat affected zone of the weld. The fracture surfaceswere too abraded to positively identify the initiation point; small cracks, however, werealso identified within the weld under microscopic examination. The original flaw whicheventually propagated and produced the rupture, apparently in this case, just happenedto be near the weld rather than in the weld